Imaging strategists ply industry with new ideas

High-temp superconductors, implantable ultrasound, 256-slice CT systems pledge to remake landscape of medical imaging

On the wings of new technologies, vendors have achieved record sales thanks to such innovations as 3T MR, 64-slice CT, and fast PET/CT hybrids. But turbulence lies ahead.

Imaging's growing popularity has attracted the attention of number crunchers in the U.S. government and third-party payers who would like to reduce healthcare costs. Any industry as dependent on technology as medical imaging naturally looks for solutions in technologies to drive sales.

New clinical capabilities, lower costs of manufacture, and increased productivity by users are the guiding lights of industry strategists. Comprehensive solutions will come from well-known vendors, but some of the least known players, start-up companies only beginning to get their corporate legs, are plying the imaging community with new ideas.

MR'S MIXED BAG

Double-digit revenue growth for 3T in recent years has been offset by drops in the MR mainstays: open scanners and the current but embattled benchmark of this modality, 1.5T. In 2005, U.S. customers spent $1.35 billion on new MR scanners compared with $1.5 billion the year before. These totals compare with $1.4 billion in 2003 and $1.5 billion in 2002. This year, the market is likely to support $1.4 billion in shipped scanners.

Providers seeking an edge on competitors while getting ready for the expected boom in clinical applications made possible by 3T's stronger signal have been turning to 3T MR, which vendors are featuring as the new standard in high-field imaging. But interest in 3T is beginning to ebb, as the early adopters of this technology are sated and continued growth depends on demand from the mainstream imaging community.

This reaction to 3T is happening just as the community appears to have lost its commitment to open scanners. Last year's fall in revenue for new scanners shipped to U.S. sites was largely due to a huge drop in the market for open scanners, which saw unit numbers fall 38% and dollar value tumble 62% to under $160 million, according to consolidated industry estimates. This is less than the revenue generated by 3T sales.

Battling this trend are Philips' Panorama Open 1.0T system and Siemens' wide-bore Espree 1.5T, each of which has engineered solutions that balance image quality and patient comfort. Future developments in open scanning, however, may hinge on the development of core technologies, particularly high-temperature superconductors that will allow innovative designs and lower production costs.

Supercon, a manufacturer of superconducting wire and cable in Shrewsbury, MA, is trying to develop the high-temperature superconductor needed to operate an open MR scanner at 1.5T and above. The metal involved would superconduct at 20 degrees higher than absolute zero, extraordinarily cold by most standards but warm enough to allow the magnet to be cooled using a mechanical cryocooler rather than cryogens. This would eliminate the vacuum that must otherwise be created, as well as the space necessary for a cryostat, opening the door to the development of the "split pair" coil configuration needed for an open system MR.

Such innovations may take MR into niches created specifically by clinical needs. One would lead to the development of a dedicated neuro scanner. Dave Ferguson, general manager of core products for GE's MR business, foresees the time when a 3T magnet can be easily and cost-effectively fabricated in a configuration small enough to fit in a nontraditional space.

"We're not talking about a schoolbus-size 3T with a tiny field-of-view," he said. "It would have to be a small form factor magnet that could be sited where 3Ts cannot be sited today."

The demand for such a device would come from neurologists in response to a rising incidence of neurodegenerative disease. But it would come only with a breakthrough in high-temperature superconductors.

"One of the things that make magnets as large as they are is the amount of cryogens needed to cool them," he said.

Other ways exist to bring MR to small spaces. Engineers at Advanced Imaging Research in Cleveland have fitted radiofrequency coils into neonatal incubators. These coils, which interface with 1.5T scanners, allow high-resolution imaging of the brain, spine, and heart. Head and body coils have been built into an incubator along with a ventilator, vital signs monitor, and infusion pumps-all compatible with MRI and spectroscopy. The incubator is mounted on a trolley and features battery backup to run and monitor life support, permitting transport from the neonatal intensive care unit to anywhere in the hospital. It works with GE, Siemens, and Philips 1.5T systems. Work is under way to develop similar products for 0.7T and 3T scanners.

BORN-AGAIN CTS

A surge in the adoption of 64-slice CT scanners last year is showing signs of slowing, and the near term is uncertain, although the industry is still reveling in the wake of the best year on record. Revenue from the delivery of new CT units in the U.S. reached $1.5 billion, a 15% leap beyond the previous year on a 3% growth in unit volume. The availability of 64-slice scanners capable of coronary CT angiography led the industry to those heights.

Initial adoption has been by radiologists who have applied these scanners to conventional applications while dabbling in cardiology. The second stage of the sales trajectory of 64-slice scanners is expected to come from cardiologists seeking a diagnostic answer to coronary imaging, which likely will not appear for another 18 to 24 months. By then, the landscape of CT could be in for a shock.

A prototype 256-slice scanner is already generating 4D images in Japan, and a commercial version could be available in two years. Other manufacturers are known to be experimenting with similar prototypes, as well as ones constructed from flat panels.

The Japanese prototype, the result of a collaboration between Toshiba and the government of Japan, has captured dynamic volumetric images of the liver and other tissues. Such CTs could potentially be used to produce dynamic studies of the brain as well as the renal and, possibly most important, coronary arteries.

"When you can image an organ in a single rotation, you are no longer tied to the boundaries of helical scanning," said John Zimmer, vice president of marketing for Toshiba America Medical Systems. "When imaging the heart, you don't have to acquire data over several heartbeats. You acquire the entire organ at the same time."

Trying to keep up with the data streaming from such a scanner, however, could be a nightmare. Data acquired using Toshiba's prototype 256-slice scanner are transferred from the gantry using laser diode and photodiode pairs at a net transfer speed of 5 Gbps. This capability only hints at the challenges to be addressed in image reconstruction.

"The goal is not to see slices but a 3D volume immediately," Zimmer said.

Data handling can be an opportunity as well as a barrier, depending on your perspective. InstaRecon is looking at the use of iterative algorithms that promise high-quality images from low-dose data. Engineers at the Urbana, IL, company are building detailed models of the physics and statistics of the data acquisition process, so as to amplify the content of data mathematically and, thereby, reduce the need to collect so much. Reducing data needs means giving patients lower doses.

The algorithms for doing so exist already, but they are time-consuming, raising the specter of another bottleneck in an already time-constrained process. But fast algorithms, developed and patented by the University of Illinois, for the backprojection and reprojection steps needed for these iterative algorithms, speed up the processing by 10 to 50 times.

Backprojection is the assignment of attenuation values to individual pixels along the path of an x-ray as it travels through the patient. Reprojection is the process of simulating what the scanner does, so as to ensure that the values determined by backprojection are correct. In CT, reprojection is used to correct flaws in the data, such as beam-hardening artifacts.

InstaRecon hopes to use its accelerated iterative algorithms to create reduced-dose, high-precision CT. In their most advanced form, these algorithms may be applicable not only to the multislice scanners of today but the flat-panel systems yet to come.

"If you can afford to do more computing, you can get by with lower x-ray dose," said Yoram Bresler, president, chief technology officer, and cofounder of InstaRecon.

The acceleration possible with such algorithms translates into lower requirements on the computing hardware necessary to achieve a given resolution or speed of reconstruction. The need for faster processing is apparent already with 64-slice CTs. It will grow more acute with advanced systems such as Siemens Somatom Definition, which features two separate x-ray imaging chains-two x-ray tubes and two detectors-in a single machine. The idea behind the Definition is to improve temporal resolution, but the ability to operate two x-ray beams at different energies could yield other benefits.

Dual-energy scanning might be used in cardiovascular imaging to differentiate the vessel wall from surrounding tissue. It could be used to isolate bone and blood vessels, as in the case of complex vasculature near the skull or across the knee. It may also be used in the lung to characterize pulmonary arteries and nodules or in the liver to differentiate fatty lesions from carcinoma. Digital subtraction, with one beam tuned to a contrast medium and the other to general tissue, might be accomplished in a snap.

"By beginning to look at some of those applications, you can see how we can reduce the amount of time the patient needs to be on the table or how we can reduce the number of steps after the acquisition," said Joe Camaratta, vice president of Siemens global solutions for the U.S.

GE and Philips expect to achieve the same clinical effect with a single beam, albeit with some fancy computation. One technique both companies are working on, called spectral CT, translates photon strikes into anatomic position and material or tissue characteristics. Among the possibilities might be distinguishing contrast media, calcifications, and bone versus soft tissue, particularly blood vessel walls. GE has a detector capable of spectral imaging in prototype and has generated images using phantoms.

"Ultimately, we want to use the same x-ray beam, discerning from it what the (body) materials are, in a way that doesn't overexpose patients to dose," said Gene Saragnese, global vice president of molecular imaging and CT at GE Healthcare. "We are looking at a technique by which we count the photons and characterize their energies as they come in."

Because the signal contains so much information, this particular technique prefers low dose.

"It doesn't work very well if you slam the detector with the kind of flux levels seen in today's CT scanners," said Brian J. Duchinsky, global general manager for GE Healthcare's CT business.

To help process these data, vendors hope to increase the digitalization of detection systems, miniaturizing the signal processing electronics by making them part of the detector module itself.

"This will allow faster acquisition times, more channels, and much better, cleaner signals," said Diego Olego, chief technology officer for Philips Medical Systems.

CT scanners won't necessarily get smaller. The miniaturization will make room for the bigger detectors that are all but inevitable, he said.

PET/CT'S AFFAIR OF THE HEART

Despite overcapacity in the imaging community, characterized by hybrid scanners sitting idle much of the day, PET/CT has attracted a growing crowd of purchasers made up for the most part of imaging facilities seeking to keep up with one another. PET/CT sales last year jumped 23% in new unit shipments in the U.S., generating revenues of $390 million in 2005 compared with $315 million in 2004 and $310 million in 2003.

Most sales have been for oncological applications, raising questions about how many more sites can be supported by the current referral base. But new growth may be on the horizon. The first commercial versions of 64-slice PET/CTs dedicated to cardiac applications have begun to appear.

GE's Discovery VCT and Siemens' Biograph 64 combine 64-slice CT and high-resolution PET as a one-stop shop for assessment of suspected cardiovascular disease. The PET component metabolically maps the heart, while the CT supports coronary angiography, requiring a breath-hold only several seconds long.

Combining the two modalities and conducting simultaneous exams gets rid of the misalignment that can result when PET and CT images are taken separately. But rest and stress imaging using the Discovery VCT to identify perfusion defects presents special challenges that require the deft use of CT during pharmacolgical stress, which causes the heart and diaphragm to shift upon injection of the agent, and a special algorithm developed by GE.

"Our 'shift algorithm' is specifically designed to pull up the attenuation correction and to match the CT and the PET perfusion to see if they line up and, if not, actually line them up and reconstruct them," said Karthik Kuppusamy, Ph.D., GE's general manager for nuclear medicine, PET/CT, and cyclotron business.

Meanwhile, vendors are hoping to appeal to those who want the very best, offering PET/CTs that go faster and deliver images with better quality than ever before. Siemens' Biograph TruePoint PET/CT and Philips' Gemini TF (TrueFlight) promise to dramatically cut the time needed to perform PET exams.

Siemens' decision to add crystal to build out the detector ring on its Biograph PET/CT scanner increases axial coverage by 33%, according to the company. The result is a jump in photon detection of 78%, which translates into faster scan times or higher quality images.

ULTRASOUND GETS SMALL

Despite a major turn toward volumetric imaging and the adoption of other high technologies, the ultrasound market acts as if sonographic products were commodities. Prices are moderated by strong competition, and sales are ruled mostly by demand for replacement systems. The ultrasound market for new unit sales in the U.S. last year edged upward to about $870 million from $820 million the year before. It failed to reach the revenue mark of just over $1 billion set in 2003, however. The cardiology segment jumped 15% to about $300 million from $260 million in 2005. Radiology was virtually flat compared with the previous year at $380 million. Ob/gyn rose to $136 million from $129 million.

Buried in this heap of market mediocrity, however, is a star performer: hand-carried ultrasound. SonoSite once had this marketplace largely to itself. GE has made it a priority with the introduction this spring of laptop-sized ultrasound systems that address opportunities in radiology, ob/gyn, the emergency and operating rooms, echo labs, and physician offices. Look for Philips and Siemens to follow suit.

Continued miniaturization promises to take ultrasound into new and challenging places. Nanotechnology called cMUT (capacitive microfabricated ultrasonic transducers) is leading GE Healthcare to develop a 3D ultrasound system that can be mounted in an endoscope to enable image-guided cancer surgery. Miniaturized 2D and 3D ultrasound imaging devices, operating at 10 to 15 MHz, are also on the company's drawing board for integration with electrode and RF ablation devices.

The goal is to monitor the ablation procedure by visualizing the lesion and mapping the distribution of temperature during RF delivery. The devices would also assess the heart before, during, and after surgery.

Ultrasound technology might even be embedded into artificial vascular grafts. About 60,000 of these grafts are performed in the U.S. each year. They fail when blood flow diminishes. Smart grafts embedded with Doppler sensors would measure blood flow, transmitting the data wirelessly to an external monitor that would set off an alarm if flow drops.

Mr. Freiherr is business editor of Diagnostic Imaging.